[Haskell-cafe] Finally tagless and ghc-extensions

[Günther Schmidt]

Hi all,

ok, I haven't let go on the EDSL subject just yet. I'm still digesting the  
"Finally Tagless" approach from Oleg, Jacques and Chen.

In both, the JFP & APLAS paper, the authors note that no fancy type  
extension are needed for this approach. But in almost every (Haskell)  
example that is a bit more involved the authors do make use of functional  
dependencies and / or type families. GADTs are also being used, although  
*not* in the syntax definition but merely in the semantic instances, I'm  
cool with that :) .

I haven't touched base yet with either functional dependencies nor type  
families as I understand the edges on neither are yet worked out, so I'd  
prefer being able to avoid either when using the tagless approach even in  
more involved settings.

Is that possible?


Günther